INDEX

Explanations

maps

oai_token-act-pair · gpt-3.5-turbo

references to maps

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

nuts

-0.83

nder

-0.77

icer

-0.76

ruciating

-0.75

vae

-0.75

ntil

-0.71

idth

-0.67

aneous

-0.67

ente

-0.66

ership

-0.66

POSITIVE LOGITS

ographer

0.88

map

0.87

 mapped

0.78

 sheet

0.78

 maps

0.77

 outline

0.76

 overlay

0.74

map

0.73

MAP

0.72

MAP

0.72

Activations Density 0.011%

maps

references to maps

No Comments

No Known Activations