INDEX

Explanations

specific acronyms

oai_token-act-pair · gpt-3.5-turbo

references to specific types of medical or technical classifications

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

fulness

-0.81

milo

-0.74

ufact

-0.74

coni

-0.70

éĹĺ

-0.69

grass

-0.68

 Magicka

-0.68

 fullest

-0.67

âĸ¬

-0.67

fitting

-0.65

POSITIVE LOGITS

UE

1.05

atari

0.95

ues

0.91

WER

0.89

addafi

0.89

atana

0.86

ued

0.86

uably

0.83

DN

0.82

atar

0.81

Activations Density 0.011%

specific acronyms

references to specific types of medical or technical classifications

No Comments

No Known Activations