INDEX

Explanations

mentions of the word "shot"

oai_token-act-pair · gpt-3.5-turbo

references to various "shots" in a cinematic or photographic context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

herty

-0.78

ophers

-0.76

anwhile

-0.71

ophical

-0.70

agara

-0.68

undai

-0.66

bye

-0.65

ionage

-0.65

ungle

-0.64

itably

-0.64

POSITIVE LOGITS

 Shot

1.46

 Shooting

1.00

Shot

0.96

creen

0.89

 Shoot

0.87

 Shots

0.87

chens

0.87

 Caption

0.80

nikov

0.77

shot

0.77

Activations Density 0.007%

mentions of the word "shot"

references to various "shots" in a cinematic or photographic context

No Comments

No Known Activations