INDEX

Explanations

details related to specific dates, locations, and events

oai_token-act-pair · gpt-3.5-turbo

references to significant events and dates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alam

-0.64

Reviewer

-0.62

zech

-0.59

phas

-0.58

utations

-0.56

Score

-0.56

akedown

-0.55

\)

-0.55

ride

-0.55

hett

-0.54

POSITIVE LOGITS

 February

1.75

 July

1.75

 October

1.74

 April

1.73

 September

1.73

 March

1.71

 December

1.71

 November

1.70

Feb

1.70

 June

1.69

Activations Density 1.557%

details related to specific dates, locations, and events

references to significant events and dates

No Comments

No Known Activations