INDEX

Explanations

strongly opinionated statements or beliefs

oai_token-act-pair · gpt-3.5-turbo

references to beliefs or ideologies that are framed as problematic or dangerous

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

achu

-0.76

Edit

-0.74

oop

-0.73

aido

-0.72

options

-0.71

acid

-0.71

ents

-0.70

ecided

-0.69

intent

-0.67

otte

-0.67

POSITIVE LOGITS

 manifestation

1.24

 reflection

1.20

 symptom

1.15

 reminder

1.12

 testament

1.07

 continuation

1.02

 culmination

1.02

 distraction

1.01

 betrayal

1.00

 contradiction

0.99

Activations Density 0.171%

strongly opinionated statements or beliefs

references to beliefs or ideologies that are framed as problematic or dangerous

No Comments

No Known Activations