INDEX

Explanations

details related to actions and events

oai_token-act-pair · gpt-3.5-turbo

references to involvement or participation in actions or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Funk

-0.64

 Millennium

-0.64

 Berk

-0.60

 Chancellor

-0.59

 Dunk

-0.58

 Kramer

-0.57

 Judd

-0.57

Emb

-0.55

 Moose

-0.54

 Advance

-0.54

POSITIVE LOGITS

etheless

1.08

bda

0.72

conservancy

0.71

arthed

0.66

urches

0.64

prus

0.63

dating

0.63

staking

0.63

entimes

0.63

rael

0.63

Activations Density 0.863%

details related to actions and events

references to involvement or participation in actions or events

No Comments

No Known Activations