INDEX

Explanations

references to "SM" followed by a numerical value, such as "SM 10" or "SM 9"

oai_token-act-pair · gpt-3.5-turbo

mentions of "SM" or related abbreviations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ãĥĥãĥĪ

-0.74

 Frie

-0.73

Ved

-0.73

ãĤ¡

-0.72

 Gamergate

-0.71

wright

-0.71

ãĥĥãĥī

-0.68

icular

-0.68

 Lyme

-0.68

reon

-0.67

POSITIVE LOGITS

ooth

1.05

ugg

0.99

ITH

0.96

ART

0.91

iley

0.91

ould

0.90

achine

0.90

ASH

0.87

oke

0.84

oby

0.84

Activations Density 0.011%

references to "SM" followed by a numerical value, such as "SM 10" or "SM 9"

mentions of "SM" or related abbreviations

No Comments

No Known Activations