INDEX

Explanations

phrases related to enforcing rules, mandates, or duties

oai_token-act-pair · gpt-3.5-turbo

mentions of authority and imposition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Tribe

-0.62

 voyage

-0.58

beck

-0.58

 Sioux

-0.57

 Clyde

-0.56

 room

-0.56

ovies

-0.56

 Bethlehem

-0.56

kers

-0.56

adelphia

-0.55

POSITIVE LOGITS

ocide

0.71

 mantra

0.71

 stricter

0.70

WS

0.70

onz

0.69

 compulsion

0.68

rules

0.68

azon

0.67

bug

0.67

atars

0.67

Activations Density 0.261%

phrases related to enforcing rules, mandates, or duties

mentions of authority and imposition

No Comments

No Known Activations