INDEX

Explanations

terms related to deterrence and discouragement

oai_token-act-pair · gpt-3.5-turbo

terms related to deterrence and discouragement

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

 Puppet

-0.68

 Awesome

-0.65

vo

-0.65

ologue

-0.63

 Open

-0.63

 Expression

-0.63

 nods

-0.63

byn

-0.62

asia

-0.62

 Burk

-0.61

POSITIVE LOGITS

 deterrent

3.45

 deterrence

2.65

 dissu

2.43

 discourage

2.34

 deter

2.30

 discouraged

1.90

 discour

1.84

 discouraging

1.80

 Deter

1.44

 intimidate

1.13

Activations Density 0.035%

terms related to deterrence and discouragement

terms related to deterrence and discouragement

No Comments

No Known Activations