INDEX

Explanations

references to attacks or attackers

oai_token-act-pair · gpt-3.5-turbo

references to an attacker in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

zl

-0.95

wheel

-0.77

sit

-0.76

umph

-0.68

carb

-0.68

spring

-0.66

REE

-0.65

 Hemp

-0.64

 Cooking

-0.63

 Mineral

-0.63

POSITIVE LOGITS

 attacker

1.01

 attackers

0.96

 attacked

0.84

 attack

0.79

 assailant

0.79

 assailants

0.78

attack

0.72

CVE

0.69

Sof

0.68

idis

0.68

Activations Density 0.011%

references to attacks or attackers

references to an attacker in various contexts

No Comments

No Known Activations