INDEX

Explanations

mentions of criminal offenders

oai_token-act-pair · gpt-3.5-turbo

references to offenders in a criminal context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

quickShipAvailable

-0.76

printed

-0.75

zl

-0.72

porting

-0.71

planes

-0.70

lad

-0.70

upp

-0.67

HAHA

-0.66

bind

-0.66

ARY

-0.66

POSITIVE LOGITS

 offender

1.04

 offenders

1.02

 ineligible

0.83

 offending

0.82

 trespass

0.82

 offend

0.80

 sentenced

0.79

 misuse

0.77

 pled

0.76

 tresp

0.76

Activations Density 0.011%

mentions of criminal offenders

references to offenders in a criminal context

No Comments

No Known Activations