INDEX

Explanations

instances of retaliatory actions or motives

oai_token-act-pair · gpt-3.5-turbo

terms associated with retaliation and responses to provocation

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

âĢ¢âĢ¢

-0.76

opathy

-0.74

 Matter

-0.74

thus

-0.72

 Dying

-0.69

corn

-0.68

than

-0.67

opic

-0.65

esan

-0.65

oret

-0.65

POSITIVE LOGITS

 retaliation

1.31

 retali

1.28

 retaliate

1.28

 retribution

1.02

 repr

1.00

 punishment

0.87

 against

0.83

 punitive

0.83

 deterrent

0.80

 enthusi

0.80

Activations Density 0.006%

instances of retaliatory actions or motives

terms associated with retaliation and responses to provocation

No Comments

No Known Activations