INDEX

Explanations

words related to preliminary or initial information

oai_token-act-pair · gpt-3.5-turbo

words related to relationships and connections between individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

çīĪ

-0.68

 Schwar

-0.67

 Schr

-0.63

catentry

-0.62

OWS

-0.56

ALLY

-0.56

OLD

-0.56

enegger

-0.55

HQ

-0.55

çĶ°

-0.54

POSITIVE LOGITS

iminary

0.86

otal

0.77

emic

0.75

umatic

0.74

asant

0.73

emouth

0.73

odon

0.71

iasis

0.70

ient

0.68

eline

0.68

Activations Density 0.103%

words related to preliminary or initial information

words related to relationships and connections between individuals

No Comments

No Known Activations