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

umbai

-0.77

NOW

-0.69

 fortnight

-0.67

 fuelled

-0.66

ibaba

-0.65

ibal

-0.65

 Flood

-0.64

 lakh

-0.62

beer

-0.61

vernment

-0.60

POSITIVE LOGITS

sth

0.70

ients

0.67

 Scha

0.66

ĺħ

0.64

LV

0.64

 handcuffs

0.63

tc

0.62

SQ

0.61

efe

0.60

ts

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.