INDEX

Explanations

instances of the word "sting" related to undercover operations or deceitful activities

oai_token-act-pair · gpt-3.5-turbo

references to law enforcement operations and undercover activities

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

 Arbor

-0.85

McM

-0.82

rÃ©

-0.80

np

-0.70

ilities

-0.69

books

-0.68

rail

-0.65

deg

-0.65

issues

-0.64

 cereal

-0.62

POSITIVE LOGITS

 sting

4.04

 Sting

2.27

 undercover

1.84

 bait

1.34

 informant

1.23

 informants

1.11

 stink

1.07

 venom

0.99

 extortion

0.95

 bounty

0.94

Activations Density 0.037%

instances of the word "sting" related to undercover operations or deceitful activities

references to law enforcement operations and undercover activities

No Comments

No Known Activations