INDEX

Explanations

business-related terms or entities

oai_token-act-pair · gpt-3.5-turbo

names of organizations, clubs, or programs

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

 practicable

-0.72

 externalToEVAOnly

-0.63

wikipedia

-0.61

inventoryQuantity

-0.60

 acknow

-0.60

aternal

-0.59

utory

-0.59

SPONSORED

-0.58

ItemImage

-0.58

icable

-0.58

POSITIVE LOGITS

LLC

0.91

Ltd

0.89

 Magazine

0.88

 Presents

0.87

HQ

0.84

âĦ¢

0.83

rior

0.83

riors

0.81

ners

0.80

 Finder

0.79

Activations Density 0.388%

business-related terms or entities

names of organizations, clubs, or programs

No Comments

No Known Activations