INDEX

Explanations

names of companies or organizations

oai_token-act-pair · gpt-3.5-turbo

references to the company "Bio" along with related terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

UAL

-0.93

sub

-0.80

eer

-0.75

adoes

-0.75

anguage

-0.73

 tune

-0.67

mallow

-0.65

otor

-0.65

âĶĢâĶĢ

-0.64

awaru

-0.64

POSITIVE LOGITS

Bio

1.09

Bio

0.96

osate

0.90

 Asylum

0.85

acter

0.81

 Roche

0.80

zyme

0.78

 Resp

0.75

 Kurd

0.73

 Industries

0.73

Activations Density 0.008%

names of companies or organizations

references to the company "Bio" along with related terminology

No Comments

No Known Activations