INDEX

Explanations

terms related to abstract concepts or beliefs, especially those involving strong opinions or ideologies

oai_token-act-pair · gpt-3.5-turbo

instances of the term "doctrine" and related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ells

-0.77

vals

-0.67

ALTH

-0.64

cknowled

-0.64

taining

-0.63

bors

-0.62

 Canary

-0.62

 Democr

-0.62

azar

-0.62

Interstitial

-0.62

POSITIVE LOGITS

 doctrine

1.28

 doctrines

0.96

 Doctrine

0.92

 guiActiveUn

0.87

istries

0.82

ually

0.81

ologies

0.74

arium

0.71

 stance

0.70

 droid

0.69

Activations Density 0.008%

terms related to abstract concepts or beliefs, especially those involving strong opinions or ideologies

instances of the term "doctrine" and related concepts

No Comments

No Known Activations