INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 slab

-0.63

LM

-0.62

Īè

-0.62

cp

-0.61

 premise

-0.60

thw

-0.59

 baker

-0.59

 outing

-0.58

 chim

-0.58

 Penguin

-0.58

POSITIVE LOGITS

cone

0.77

ait

0.75

aments

0.75

isco

0.73

upe

0.73

icter

0.72

ivities

0.71

Ur

0.70

lie

0.70

eret

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.