INDEX

Explanations

mentions of values, especially in the context of principles, ethics, and belief systems

oai_token-act-pair · gpt-3.5-turbo

references to ethical or moral values

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

å¥³

-0.79

geon

-0.73

igans

-0.71

waves

-0.69

sie

-0.68

bare

-0.68

rontal

-0.68

jen

-0.67

wa

-0.64

Sus

-0.63

POSITIVE LOGITS

 ideals

0.77

 values

0.77

 tolerance

0.75

iblings

0.74

 principles

0.71

 proposition

0.69

 guiding

0.68

 beliefs

0.68

 sensibilities

0.68

 propositions

0.66

Activations Density 0.013%

mentions of values, especially in the context of principles, ethics, and belief systems

references to ethical or moral values

No Comments

No Known Activations