INDEX

Explanations

expressions or mentions related to moral principles or beliefs

oai_token-act-pair · gpt-3.5-turbo

mentions of abstract principles or moral standards

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

geon

-0.77

brain

-0.74

--------------------------------------------------------

-0.72

rontal

-0.71

Runner

-0.69

sie

-0.68

interrupted

-0.68

waves

-0.67

slow

-0.67

wards

-0.66

POSITIVE LOGITS

 Values

0.92

 values

0.90

values

0.75

 ideals

0.74

Values

0.73

 embodied

0.73

 proposition

0.72

zsche

0.71

 value

0.69

 propositions

0.67

Activations Density 0.013%

expressions or mentions related to moral principles or beliefs

mentions of abstract principles or moral standards

No Comments

No Known Activations