INDEX

Explanations

mathematical or programming content, particularly symbols and commands

oai_token-act-pair · gpt-3.5-turbo

symbols, numbers, and references related to data or citations

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

Ħ¢

-0.69

 paran

-0.62

 Monstrous

-0.57

 challeng

-0.57

 Alto

-0.56

 empt

-0.55

hement

-0.54

Ń·

-0.54

 fool

-0.54

 facilit

-0.53

POSITIVE LOGITS

soDeliveryDate

0.95

0.81

nance

0.77

ochond

0.72

tumblr

0.70

FORMATION

0.66

mosp

0.65

aughs

0.65

itars

0.64

ril

0.63

Activations Density 0.141%

mathematical or programming content, particularly symbols and commands

symbols, numbers, and references related to data or citations

No Comments

No Known Activations