INDEX

Explanations

references to actions that users can perform or services they can access

oai_token-act-pair · gpt-3.5-turbo

modal verbs indicating capability or permission

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ONSORED

-0.76

soType

-0.75

 Execution

-0.66

 Mant

-0.65

 reven

-0.64

 Means

-0.63

interstitial

-0.63

fixed

-0.62

soDeliveryDate

-0.62

killing

-0.60

POSITIVE LOGITS

't

1.14

 choose

1.13

 participate

1.09

 enjoy

1.08

 afford

1.07

 indulge

1.05

 rejoice

1.05

 attest

1.04

 easily

1.03

 learn

1.03

Activations Density 0.147%

references to actions that users can perform or services they can access

modal verbs indicating capability or permission

No Comments

No Known Activations