INDEX

Explanations

references to police-related content

oai_token-act-pair · gpt-3.5-turbo

references to police activities or mentions of police

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

gets

-0.78

ocene

-0.74

uces

-0.73

cast

-0.71

righteous

-0.71

resents

-0.71

arial

-0.70

isable

-0.68

vid

-0.67

comes

-0.67

POSITIVE LOGITS

 officers

1.22

 allege

1.06

 Officers

1.05

 spokeswoman

1.04

 detectives

1.04

say

1.04

 Superintendent

1.02

 spokesman

1.02

 officer

1.01

Sgt

1.00

Activations Density 0.059%

references to police-related content

references to police activities or mentions of police

No Comments

No Known Activations