INDEX

Explanations

phrases or words related to risky actions or decisions

oai_token-act-pair · gpt-3.5-turbo

references to risky behavior or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Nap

-0.89

anooga

-0.86

yers

-0.80

upon

-0.77

spons

-0.76

ann

-0.75

hun

-0.72

bert

-0.72

frey

-0.72

hunt

-0.71

POSITIVE LOGITS

 risky

1.26

 gamble

1.22

 gamb

1.02

 bets

0.94

 proposition

0.83

 risks

0.82

 maneuver

0.81

 risk

0.80

 maneuvers

0.78

ModLoader

0.76

Activations Density 0.019%

phrases or words related to risky actions or decisions

references to risky behavior or actions

No Comments

No Known Activations