INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

osate

-0.75

qus

-0.74

wana

-0.70

yip

-0.69

âķĲ

-0.69

ricted

-0.69

agent

-0.68

appa

-0.66

venants

-0.66

olate

-0.65

POSITIVE LOGITS

theless

0.70

 evils

0.69

 partisans

0.69

illac

0.66

 unfocusedRange

0.65

 highs

0.64

Pope

0.62

tics

0.62

Vit

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.