INDEX

Explanations

phrases indicating a conclusion or a statement that serves as an introduction to an opinion or information

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a transition or introduction of new information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

aneous

-0.67

umb

-0.60

aneously

-0.58

 Dragonbound

-0.58

kefeller

-0.56

EMA

-0.54

link

-0.54

 slashing

-0.54

 runoff

-0.53

assad

-0.53

POSITIVE LOGITS

said

0.85

 said

0.84

Ĥ¬

0.70

pects

0.69

©¶æ

0.68

Ģ

0.65

hemoth

0.64

Â¯

0.64

 conceded

0.62

Ł

0.62

Activations Density 0.049%

phrases indicating a conclusion or a statement that serves as an introduction to an opinion or information

phrases that indicate a transition or introduction of new information

No Comments

No Known Activations