INDEX

Explanations

numerical values mixed with words related to computers and technology

oai_token-act-pair · gpt-3.5-turbo

numerical values and statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 compr

-0.79

uyomi

-0.76

matically

-0.70

zag

-0.70

wich

-0.69

acco

-0.69

falls

-0.67

uden

-0.66

opter

-0.63

unks

-0.62

POSITIVE LOGITS

çĶ°

0.86

Ñĭ

0.84

ropolitan

0.82

ASC

0.74

REM

0.73

0.72

%:

0.70

rpm

0.70

 cents

0.69

Activations Density 0.065%

numerical values mixed with words related to computers and technology

numerical values and statistics

No Comments

No Known Activations