INDEX

Explanations

reports of violent incidents and conflicts in various regions or countries

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Reason

-0.89

acea

-0.82

Truth

-0.73

worthiness

-0.73

vana

-0.73

Guide

-0.71

etus

-0.70

Marvel

-0.70

 Wilde

-0.70

 Entered

-0.70

POSITIVE LOGITS

 clashes

1.69

 bombings

1.54

 bloodshed

1.46

 attacks

1.39

 killings

1.39

 shelling

1.39

 airstrikes

1.37

 massacres

1.37

 raids

1.35

 riots

1.29

Activations Density 0.228%

reports of violent incidents and conflicts in various regions or countries

No Comments

No Known Activations