INDEX

Explanations

violent actions or descriptions, including tearing, ripping, and chopping

oai_token-act-pair · gpt-3.5-turbo

violent or destructive actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

hof

-0.67

rophe

-0.64

lain

-0.64

 Lobby

-0.63

 Colomb

-0.60

lied

-0.60

MG

-0.60

Support

-0.60

ITNESS

-0.60

abilities

-0.59

POSITIVE LOGITS

 away

1.03

 apart

1.02

 bones

0.96

awed

0.93

awa

0.92

 down

0.90

 nails

0.87

adoes

0.85

 throats

0.84

 tires

0.82

Activations Density 0.162%

violent actions or descriptions, including tearing, ripping, and chopping

violent or destructive actions

No Comments

No Known Activations