INDEX

Explanations

descriptions or technical information within a text

oai_token-act-pair · gpt-3.5-turbo

textual elements that signify content organization or structure

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pled

-0.73

 indistinguishable

-0.63

ilaterally

-0.63

 pierced

-0.62

ilater

-0.62

anwhile

-0.61

 diseng

-0.61

 foremost

-0.60

 thereafter

-0.60

 symp

-0.60

POSITIVE LOGITS

 Transcript

1.12

 Description

0.97

 Summary

0.96

][

0.90

 Definitions

0.90

 Finder

0.89

 Recipe

0.87

Description

0.87

 Pages

0.87

 Introduction

0.87

Activations Density 0.134%

descriptions or technical information within a text

textual elements that signify content organization or structure

No Comments

No Known Activations