INDEX

Explanations

names of places and people

oai_token-act-pair · gpt-3.5-turbo

references to data and identifiers related to various organizations and entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

oppable

-0.70

ishers

-0.68

isters

-0.68

 compe

-0.64

bledon

-0.63

umenthal

-0.63

 steady

-0.63

owler

-0.63

anchester

-0.62

ITNESS

-0.61

POSITIVE LOGITS

ensis

0.84

 Parish

0.83

 Potion

0.80

 Sauce

0.79

 Experimental

0.75

 Tribal

0.75

 Islands

0.75

 Tyrann

0.74

 Unified

0.74

 Meditation

0.74

Activations Density 0.313%

names of places and people

references to data and identifiers related to various organizations and entities

No Comments

No Known Activations