INDEX

Explanations

references to reactions and opinions related to news or events

oai_token-act-pair · gpt-3.5-turbo

mentions of reports, decisions, or actions related to jobs and courts

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

roleum

-0.85

Favorite

-0.79

utenant

-0.77

 Empires

-0.74

[_

-0.73

TAIN

-0.72

Contents

-0.72

ĩ

-0.72

ģĸ

-0.71

ategory

-0.71

POSITIVE LOGITS

 announcement

1.47

 setback

1.26

 decision

1.21

 situation

1.20

 incident

1.18

 move

1.15

 findings

1.15

 developments

1.14

 revelation

1.14

 discrepancy

1.13

Activations Density 0.271%

references to reactions and opinions related to news or events

mentions of reports, decisions, or actions related to jobs and courts

No Comments

No Known Activations