INDEX

Explanations

strings of symbols or characters, potentially indicating a specific code or language

oai_token-act-pair · gpt-3.5-turbo

instances of rare or unique terminology and non-standard formats

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

hement

-0.93

xual

-0.83

 scrap

-0.82

 realised

-0.69

 doct

-0.67

ĸļ

-0.66

 behaviour

-0.63

 milit

-0.63

illance

-0.63

 unch

-0.62

POSITIVE LOGITS

Synopsis

0.92

Chair

0.82

Reviewed

0.81

OVA

0.77

Birth

0.77

pmwiki

0.77

Includes

0.76

Discuss

0.73

Recommended

0.71

Premium

0.70

Activations Density 0.357%

strings of symbols or characters, potentially indicating a specific code or language

instances of rare or unique terminology and non-standard formats

No Comments

No Known Activations