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

km

-0.75

kas

-0.68

isans

-0.64

km

-0.64

pei

-0.63

 pard

-0.61

onomy

-0.61

kb

-0.61

 Okin

-0.60

 Pesh

-0.60

POSITIVE LOGITS

earchers

0.71

quickShipAvailable

0.69

 Dragonbound

0.68

repeat

0.66

 dilig

0.64

 Finder

0.64

 retrieval

0.64

 ensu

0.63

rehend

0.63

aunder

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.