INDEX

Explanations

references to acts of violence or confrontation

oai_token-act-pair · gpt-3.5-turbo

articles and phrases related to speculating about various risky behaviors and situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Contents

-0.92

Sources

-0.81

isites

-0.79

Links

-0.74

aci

-0.74

Versions

-0.71

cliffe

-0.69

evidence

-0.69

Events

-0.68

 indications

-0.67

POSITIVE LOGITS

 bunch

1.33

guy

1.14

 shitty

1.12

 goddamn

1.07

 crappy

1.07

kid

1.03

 stranger

1.02

lot

0.95

 thousand

0.95

 freaking

0.93

Activations Density 0.355%

references to acts of violence or confrontation

articles and phrases related to speculating about various risky behaviors and situations

No Comments

No Known Activations