INDEX

Explanations

phrases related to espionage, intelligence, or investigations

oai_token-act-pair · gpt-3.5-turbo

proper nouns and references to individuals or entities, particularly those that begin with "Rog."

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

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

âĢ¢âĢ¢

-0.69

ãĥ¡

-0.64

ites

-0.63

 Âł Âł Âł Âł

-0.63

acters

-0.63

 Âł Âł Âł Âł Âł Âł Âł Âł

-0.61

uper

-0.61

 goodbye

-0.60

pora

-0.59

ICAN

-0.58

POSITIVE LOGITS

ues

1.10

uish

1.07

UE

1.04

raphic

1.02

atory

1.02

raphics

1.01

raph

0.98

uel

0.96

allery

0.89

atories

0.89

Activations Density 0.028%

phrases related to espionage, intelligence, or investigations

proper nouns and references to individuals or entities, particularly those that begin with "Rog."

No Comments

No Known Activations