INDEX

Explanations

mentions of individuals or groups responsible for committing varying acts

oai_token-act-pair · gpt-3.5-turbo

terms related to individuals who commit crimes or harmful acts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

psey

-0.83

zl

-0.66

aver

-0.65

eem

-0.65

paio

-0.65

ria

-0.65

opian

-0.64

 Plat

-0.64

ARY

-0.64

rients

-0.62

POSITIVE LOGITS

 perpetrated

1.03

 perpetrators

0.92

 perpetrator

0.90

 offenders

0.83

 offender

0.80

 abused

0.77

 abuse

0.76

 Abuse

0.75

 spree

0.73

 Victims

0.73

Activations Density 0.011%

mentions of individuals or groups responsible for committing varying acts

terms related to individuals who commit crimes or harmful acts

No Comments

No Known Activations