INDEX

Explanations

short, concise summaries or explanations

oai_token-act-pair · gpt-3.5-turbo

phrases that encapsulate summaries or main ideas

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lean

-0.74

ali

-0.70

 faint

-0.67

rising

-0.65

rie

-0.65

young

-0.63

hung

-0.62

inates

-0.62

gypt

-0.62

claimed

-0.62

POSITIVE LOGITS

 nutshell

1.28

 summar

0.88

 gist

0.85

 summarize

0.84

 outline

0.80

 Overview

0.79

rss

0.79

 overview

0.71

 summary

0.69

="#

0.68

Activations Density 0.011%

short, concise summaries or explanations

phrases that encapsulate summaries or main ideas

No Comments

No Known Activations