INDEX

Explanations

phrases related to military or terrorist attacks

oai_token-act-pair · gpt-3.5-turbo

mentions of attacks in a geopolitical context

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

soDeliveryDate

-0.89

SourceFile

-0.86

VPN

-0.79

YP

-0.77

Plex

-0.77

natureconservancy

-0.74

ACTED

-0.74

YY

-0.72

OGR

-0.71

wx

-0.70

POSITIVE LOGITS

 behalf

1.32

etime

0.90

erous

0.88

 unsuspecting

0.82

 impulse

0.81

shore

0.81

 rampage

0.76

 spree

0.76

 innocent

0.74

eness

0.73

Activations Density 0.121%

phrases related to military or terrorist attacks

mentions of attacks in a geopolitical context

No Comments

No Known Activations