INDEX

Explanations

No Explanations Found

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

ãĤ£

-0.87

ãĥķãĤ©

-0.80

 Roller

-0.63

ãĤ·ãĥ£

-0.62

boxing

-0.61

sels

-0.60

 congr

-0.60

-0.59

XT

-0.59

à¦

-0.58

POSITIVE LOGITS

ahime

0.77

merce

0.74

icularly

0.74

theless

0.72

etheless

0.71

mes

0.70

 Rosa

0.68

 Pryor

0.68

©¶æ

0.68

uph

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.