INDEX

Explanations

questions starting with "Would"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Would" and its variations in questions or hypothetical scenarios

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

 beginning

-0.71

 division

-0.61

 belonging

-0.60

 skirm

-0.60

 integral

-0.59

 gran

-0.58

 accounting

-0.57

ang

-0.57

 roots

-0.56

 aber

-0.56

POSITIVE LOGITS

 Would

3.14

Would

2.82

 Wouldn

2.35

would

2.01

 Could

1.87

Could

1.70

 Didn

1.44

 Might

1.42

 Suppose

1.41

could

1.41

Activations Density 0.015%

questions starting with "Would"

instances of the word "Would" and its variations in questions or hypothetical scenarios

No Comments

No Known Activations