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

 veto

-0.73

rament

-0.70

rals

-0.67

qv

-0.66

 Sapp

-0.64

coni

-0.63

scr

-0.62

 beams

-0.62

iege

-0.62

NDP

-0.61

POSITIVE LOGITS

ãĥīãĥ©ãĤ´ãĥ³

0.84

ILA

0.73

FORE

0.69

ãĥ¯

0.68

Magikarp

0.68

Cla

0.68

ãĥķãĤ¡

0.67

Clark

0.67

GREEN

0.67

gae

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.