INDEX

Explanations

situations where assumptions are made

oai_token-act-pair · gpt-3.5-turbo

instances of the word "assumed" and its variations in context, indicating assumptions made by characters or people

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

zens

-0.68

fest

-0.64

hing

-0.61

lem

-0.61

bing

-0.58

rain

-0.58

abouts

-0.58

Quality

-0.57

zan

-0.57

Dialogue

-0.56

POSITIVE LOGITS

 assumed

3.61

 assume

2.19

 presumed

2.12

 assumes

1.98

 assumption

1.91

 inferred

1.72

 assuming

1.61

 assum

1.59

 implied

1.55

 assumptions

1.49

Activations Density 0.009%

situations where assumptions are made

instances of the word "assumed" and its variations in context, indicating assumptions made by characters or people

No Comments

No Known Activations