INDEX

Explanations

mentions of specific names or entities

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or entities

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

imeters

-0.79

Nak

-0.79

 necklace

-0.78

icent

-0.76

 Athen

-0.74

aki

-0.73

inventoryQuantity

-0.73

Nah

-0.73

 Pants

-0.73

 Pearce

-0.71

POSITIVE LOGITS

Ro

2.40

Ro

2.29

ro

1.89

RO

1.88

RO

1.76

ro

1.74

 Robo

1.54

 roast

1.47

Kro

1.43

 roasted

1.26

Activations Density 0.239%

mentions of specific names or entities

references to specific individuals or entities

No Comments

No Known Activations