INDEX

Explanations

statistics and figures

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to polling data and survey details

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

AFTA

-0.73

achev

-0.67

ammy

-0.66

uve

-0.66

inement

-0.62

 Centauri

-0.61

 Shining

-0.60

repair

-0.58

 Eternity

-0.58

 relativity

-0.58

POSITIVE LOGITS

 telephone

1.00

 interviews

0.94

 interviewing

0.93

 surve

0.90

 Telephone

0.89

 sampling

0.87

 respondents

0.84

 interviewed

0.83

 fielded

0.83

 interviewer

0.83

Activations Density 0.113%

statistics and figures

references to polling data and survey details

No Comments

No Known Activations