INDEX

Explanations

terms related to conflicts or confrontations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "clash" and its variations, indicating conflict or confrontation

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

Footnote

-0.69

 Receiver

-0.68

 recommends

-0.66

pod

-0.65

 Hubbard

-0.64

Newsletter

-0.64

Figure

-0.63

 Habit

-0.63

Delivery

-0.63

Hol

-0.62

POSITIVE LOGITS

 clash

3.72

 clashes

2.98

 clashed

2.31

 collide

1.79

 Clash

1.79

 showdown

1.76

 conflict

1.67

 brawl

1.60

 conflicts

1.59

 confrontation

1.55

Activations Density 0.015%

terms related to conflicts or confrontations

instances of the word "clash" and its variations, indicating conflict or confrontation

No Comments

No Known Activations