INDEX

Explanations

keywords related to policies and decision-making

oai_token-act-pair · gpt-3.5-turbo

mentions of various policies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

apy

-0.82

semble

-0.77

 Mour

-0.74

htaking

-0.73

semb

-0.72

 Stain

-0.71

lighting

-0.70

 Sparkle

-0.69

 Mile

-0.68

yards

-0.67

POSITIVE LOGITS

 policy

1.29

policy

1.14

 policies

1.13

olicy

0.96

 prescriptions

0.95

 stance

0.94

 Policies

0.94

 policymakers

0.91

 enforcement

0.85

 Policy

0.84

Activations Density 0.021%

keywords related to policies and decision-making

mentions of various policies

No Comments

No Known Activations