INDEX

Explanations

the word "Cer" or variations of it

oai_token-act-pair · gpt-3.5-turbo

references to specific locations or entities associated with the term "Cer."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

draft

-0.79

matic

-0.75

bow

-0.75

nesses

-0.75

share

-0.73

edit

-0.71

ers

-0.70

ingly

-0.70

Howard

-0.69

DEP

-0.69

POSITIVE LOGITS

Cer

1.13

cer

1.07

ategory

0.96

 ceramic

0.93

 millenn

0.93

 veter

0.88

 satell

0.85

ailability

0.85

 therap

0.84

ately

0.83

Activations Density 0.007%

the word "Cer" or variations of it

references to specific locations or entities associated with the term "Cer."

No Comments

No Known Activations