INDEX

Explanations

phrases related to rules, regulations, and standards

oai_token-act-pair · gpt-3.5-turbo

mentions of guidelines or regulations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

aunt

-0.81

mong

-0.71

rey

-0.69

 SOLD

-0.69

yn

-0.67

plex

-0.66

 Joined

-0.66

sing

-0.66

vous

-0.65

que

-0.65

POSITIVE LOGITS

 guidelines

3.52

 Guidelines

2.89

 guideline

2.49

 recommendations

1.94

idelines

1.88

 regulations

1.67

 criteria

1.64

 rules

1.62

 standards

1.61

 guidance

1.60

Activations Density 0.014%

phrases related to rules, regulations, and standards

mentions of guidelines or regulations

No Comments

No Known Activations