INDEX

Explanations

mentions of serious accusations or claims

oai_token-act-pair · gpt-3.5-turbo

references to allegations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

keys

-0.85

arton

-0.83

skill

-0.75

 busiest

-0.74

aez

-0.66

ERAL

-0.66

PB

-0.66

asus

-0.66

gencies

-0.64

options

-0.64

POSITIVE LOGITS

 against

1.31

 leveled

1.23

 levied

1.22

 alleging

1.20

 allegations

1.15

against

1.15

 leve

1.13

 accusing

1.10

 allegation

1.06

 alleges

1.04

Activations Density 0.058%

mentions of serious accusations or claims

references to allegations

No Comments

No Known Activations