INDEX

Explanations

mentions of physical actions and events, especially those related to crime or instructions

oai_token-act-pair · gpt-3.5-turbo

actions related to violent incidents or criminal activities

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ó

-0.93

iably

-0.90

oubtedly

-0.90

etheless

-0.85

inarily

-0.85

ogether

-0.85

eday

-0.84

æĺ

-0.84

uably

-0.84

conserv

-0.83

POSITIVE LOGITS

 blackout

0.82

 probe

0.80

WS

0.79

EC

0.79

 unknown

0.75

 economy

0.75

HRC

0.73

 highest

0.72

 intel

0.72

 ceiling

0.72

Activations Density 0.864%

mentions of physical actions and events, especially those related to crime or instructions

actions related to violent incidents or criminal activities

No Comments

No Known Activations