INDEX

Explanations

phrases related to capability or functionality

oai_token-act-pair · gpt-3.5-turbo

phrases related to capability and potential

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

 Bans

-0.77

hesis

-0.69

udo

-0.67

govtrack

-0.61

lights

-0.61

NCT

-0.61

lass

-0.60

dad

-0.60

bush

-0.59

 Unsure

-0.59

POSITIVE LOGITS

 withstand

0.84

Æ

0.73

umbn

0.72

 delivering

0.71

amina

0.70

IENT

0.70

 appre

0.70

 discern

0.68

 inflicting

0.67

 stren

0.66

Activations Density 0.069%

phrases related to capability or functionality

phrases related to capability and potential

No Comments

No Known Activations