INDEX

Explanations

terms related to risks and resistances, such as high-risk, flood-resistant, and antibiotic-resistant

oai_token-act-pair · gpt-3.5-turbo

terms related to risk, resistance, and environmental impact

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

Negative Logits

ĸļ

-0.84

ledge

-0.80

OPA

-0.73

amera

-0.73

ISTORY

-0.72

hack

-0.71

authorized

-0.69

stab

-0.68

ruary

-0.68

resent

-0.68

POSITIVE LOGITS

 sectors

1.18

 areas

1.14

 substances

1.10

 behaviors

1.05

 behaviours

1.04

 ones

1.04

 regions

1.03

 zones

1.03

 aspects

1.03

 elements

0.99

Activations Density 0.258%

terms related to risks and resistances, such as high-risk, flood-resistant, and antibiotic-resistant

terms related to risk, resistance, and environmental impact

No Known Activations