INDEX

Explanations

words related to action and physical movement

oai_token-act-pair · gpt-3.5-turbo

instances of the word "action" and related terms

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

bey

-0.70

arton

-0.69

aye

-0.68

YL

-0.68

redit

-0.66

thening

-0.65

cott

-0.65

 Petition

-0.64

 Mead

-0.64

DAQ

-0.61

POSITIVE LOGITS

 scenes

0.91

 sequences

0.82

fighting

0.80

 flick

0.78

 hero

0.77

 unfold

0.76

movie

0.76

packed

0.75

hero

0.75

 movie

0.74

Activations Density 0.035%

words related to action and physical movement

instances of the word "action" and related terms

No Comments

No Known Activations