INDEX

Explanations

instances where individuals are described as being good or bad at certain tasks or skills

oai_token-act-pair · gpt-3.5-turbo

phrases indicating proficiency or effectiveness in a specific skill or area

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

SPONSORED

-0.76

ships

-0.69

forth

-0.66

GROUP

-0.65

REDACTED

-0.64

released

-0.62

owner

-0.62

Registered

-0.61

loo

-0.61

blade

-0.60

POSITIVE LOGITS

 predicting

1.05

 detecting

1.00

onement

0.93

roph

0.91

 solving

0.90

 preserving

0.89

 cram

0.89

 least

0.88

 estimating

0.88

 regulating

0.85

Activations Density 0.059%

instances where individuals are described as being good or bad at certain tasks or skills

phrases indicating proficiency or effectiveness in a specific skill or area

No Comments

No Known Activations