INDEX

Explanations

mentions of news events and controversies, especially those involving legal cases or political issues

oai_token-act-pair · gpt-3.5-turbo

references to victims of police-related incidents or violence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ivalry

-0.77

 guiActiveUnfocused

-0.75

 Formation

-0.72

Style

-0.70

Studio

-0.68

iasm

-0.67

 acad

-0.66

 regulates

-0.66

 unfocusedRange

-0.65

inav

-0.65

POSITIVE LOGITS

 abducted

1.02

 suffering

0.99

 leukemia

0.98

 kidnapped

0.94

 raped

0.93

 traumat

0.93

 innocent

0.92

 murdered

0.91

 psychiat

0.89

 victimized

0.88

Activations Density 0.832%

mentions of news events and controversies, especially those involving legal cases or political issues

references to victims of police-related incidents or violence

No Comments

No Known Activations