INDEX

Explanations

phrases related to arbiters and illustrators

oai_token-act-pair · gpt-3.5-turbo

words related to decision-making authorities or regulatory bodies

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

 ankles

-0.63

hay

-0.62

JC

-0.61

 unborn

-0.61

esp

-0.61

Ji

-0.61

 Middle

-0.60

 annex

-0.60

 conting

-0.59

 shore

-0.59

POSITIVE LOGITS

rator

4.44

rators

3.50

rative

2.31

rations

2.28

ration

1.99

rating

1.96

rates

1.69

rated

1.67

rity

1.56

rate

1.41

Activations Density 0.014%

phrases related to arbiters and illustrators

words related to decision-making authorities or regulatory bodies

No Comments

No Known Activations