INDEX

Explanations

specific mentions of the term "Info" in the document

oai_token-act-pair · gpt-3.5-turbo

the repeated mention of the term "Info" and variations of "information" across the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

manship

-0.85

lock

-0.75

kus

-0.71

atos

-0.70

adj

-0.65

hip

-0.65

KR

-0.65

itiz

-0.62

 Skywalker

-0.61

itus

-0.60

POSITIVE LOGITS

 Sheet

0.89

 booths

0.78

 Gathering

0.76

 booth

0.75

Wars

0.75

 Seller

0.71

ãĤ¿

0.71

 booklet

0.70

 sheets

0.69

 gathering

0.69

Activations Density 0.035%

specific mentions of the term "Info" in the document

the repeated mention of the term "Info" and variations of "information" across the document

No Comments

No Known Activations