INDEX

Explanations

references to enemy entities or actions

oai_token-act-pair · gpt-3.5-turbo

references to enemies 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 @ 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

eret

-0.83

hots

-0.81

-0.80

auntlets

-0.76

 Printed

-0.74

ara

-0.74

orrow

-0.73

ifting

-0.73

aunder

-0.73

ranging

-0.72

POSITIVE LOGITS

 enemy

1.04

 Enemy

0.94

foe

0.91

 combatants

0.89

 adversary

0.88

 takeover

0.85

enemy

0.83

 spouse

0.77

 darling

0.75

 soldier

0.74

Activations Density 0.012%

references to enemy entities or actions

references to enemies in various contexts

No Comments

No Known Activations