INDEX

Explanations

phrases that attribute information or actions to a specific source or entity

oai_token-act-pair · gpt-3.5-turbo

the phrase "according to" followed by a source or report

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

obyl

-0.79

ograp

-0.73

gage

-0.72

ãĥİ

-0.66

ãĤ¼ãĤ¦ãĤ¹

-0.65

icably

-0.65

aque

-0.64

NetMessage

-0.63

ById

-0.62

 lasts

-0.61

POSITIVE LOGITS

 sources

0.86

 Reuters

0.76

 STATS

0.72

 according

0.71

 insiders

0.70

AFP

0.70

source

0.67

 Variety

0.66

EW

0.66

 Forbes

0.65

Activations Density 0.068%

phrases that attribute information or actions to a specific source or entity

the phrase "according to" followed by a source or report

No Comments

No Known Activations