INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

WIND

-0.81

 VIDEOS

-0.74

FL

-0.74

 footsteps

-0.68

IPM

-0.67

_-

-0.67

 GEAR

-0.67

 Whale

-0.66

ICLE

-0.66

 WRITE

-0.65

POSITIVE LOGITS

pired

0.87

arag

0.84

ansk

0.82

achev

0.79

utonium

0.75

etsk

0.74

hement

0.73

agate

0.73

cius

0.73

onial

0.73

Activations Density 0.000%

No Known Activations

This feature has no known activations.