INDEX

Explanations

phrases referring to specific conditions or situations

oai_token-act-pair · gpt-3.5-turbo

references to various conditions in different contexts

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

sonian

-1.10

umbn

-0.81

yright

-0.76

ilus

-0.73

ramid

-0.70

enic

-0.69

thouse

-0.68

 Flavoring

-0.68

ritz

-0.67

arger

-0.67

POSITIVE LOGITS

 conditions

1.11

 Conditions

0.93

 deterior

0.88

pring

0.85

 worsened

0.84

 conducive

0.81

 deteriorated

0.81

 accommodations

0.80

 condition

0.79

 conditioned

0.79

Activations Density 0.020%

phrases referring to specific conditions or situations

references to various conditions in different contexts

No Comments

No Known Activations