INDEX

Explanations

terms related to allies or allegiances

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "ally" in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Dickens

-0.69

ript

-0.68

Registration

-0.67

 Crunch

-0.65

lights

-0.63

 Process

-0.63

 Prices

-0.63

 Registration

-0.63

plates

-0.63

cases

-0.63

POSITIVE LOGITS

 ally

3.87

 allies

2.53

 allied

1.93

 Ally

1.79

foe

1.76

 Allies

1.65

 alliance

1.57

 adversary

1.46

 comrade

1.42

 teammate

1.39

Activations Density 0.012%

terms related to allies or allegiances

mentions of the term "ally" in various contexts

No Comments

No Known Activations