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

ITCH

-0.73

KI

-0.66

VAL

-0.65

enough

-0.64

oys

-0.63

itaire

-0.63

 Liter

-0.62

aina

-0.62

rikes

-0.61

istics

-0.61

POSITIVE LOGITS

 Garmin

0.66

emouth

0.65

Hes

0.64

 impe

0.62

 climbers

0.61

 perpend

0.60

 stature

0.60

 Gemini

0.60

 Canon

0.59

dom

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.