INDEX

Explanations

phrases related to eligibility or meeting specific criteria

oai_token-act-pair · gpt-3.5-turbo

instances of the word "qualify" and its variations

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

Dir

-0.67

 Trojan

-0.63

antine

-0.60

 Steps

-0.59

lyn

-0.58

mouse

-0.57

von

-0.56

 Gore

-0.56

lon

-0.56

 cent

-0.56

POSITIVE LOGITS

 qualify

3.79

 qualifies

2.59

 qualifying

1.92

 qualified

1.71

 Qual

1.55

 qualification

1.54

Qual

1.52

qual

1.44

 qualifier

1.43

 eligible

1.34

Activations Density 0.007%

phrases related to eligibility or meeting specific criteria

instances of the word "qualify" and its variations

No Comments

No Known Activations