INDEX

Explanations

mentions or references to hacking activities

oai_token-act-pair · gpt-3.5-turbo

references to individuals or groups engaged in hacking activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

eele

-0.80

 Towns

-0.79

âĶĢâĶĢâĶĢâĶĢ

-0.78

âĢ¢âĢ¢

-0.73

BuyableInstoreAndOnline

-0.72

IX

-0.72

uph

-0.67

 Veter

-0.66

 Veterinary

-0.65

illard

-0.65

POSITIVE LOGITS

intosh

1.08

 hackers

1.00

 hacker

0.93

paces

0.85

aday

0.84

 hacking

0.82

 extraord

0.82

RF

0.80

netic

0.78

punk

0.74

Activations Density 0.016%

mentions or references to hacking activities

references to individuals or groups engaged in hacking activities

No Comments

No Known Activations