INDEX

Explanations

words related to violent actions or scenarios

oai_token-act-pair · gpt-3.5-turbo

quoted phrases related to various forms of death or violence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 bending

-0.85

 destro

-0.83

 confir

-0.82

pse

-0.82

 pacing

-0.80

 packing

-0.79

 raft

-0.79

 periodic

-0.75

 arri

-0.74

¥ŀ

-0.73

POSITIVE LOGITS

soDeliveryDate

1.41

BuyableInstoreAndOnline

1.13

true

1.11

classic

1.07

false

1.05

WE

1.04

WHERE

1.04

TIT

1.04

shall

1.03

nothing

1.03

Activations Density 0.115%

words related to violent actions or scenarios

quoted phrases related to various forms of death or violence

No Comments

No Known Activations