INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

auts

-0.66

KNOWN

-0.65

TTL

-0.65

 Tone

-0.63

 Ceres

-0.63

CHA

-0.62

omp

-0.62

NEO

-0.62

trip

-0.60

 MIDI

-0.59

POSITIVE LOGITS

ngth

1.04

idth

0.75

ONSORED

0.74

 pockets

0.71

lest

0.67

ogun

0.67

adies

0.65

genic

0.65

ascus

0.65

anmar

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.