INDEX

Explanations

people or entities that are either strongly associated with a particular action, or strongly reacting to a situation

oai_token-act-pair · gpt-3.5-turbo

phrases that reference individuals or groups involved in actions or events

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

SE

-0.65

oken

-0.64

EXP

-0.62

 Anything

-0.61

CLASS

-0.61

MAS

-0.60

ÃĹ

-0.60

å½

-0.60

paren

-0.60

Untitled

-0.60

POSITIVE LOGITS

 promptly

1.07

upon

1.06

 incidentally

0.98

 oversaw

0.92

 nevertheless

0.92

 proceeded

0.91

 reportedly

0.91

 ironically

0.90

 subsequently

0.90

 understandably

0.90

Activations Density 0.134%

people or entities that are either strongly associated with a particular action, or strongly reacting to a situation

phrases that reference individuals or groups involved in actions or events

No Comments

No Known Activations