INDEX

Explanations

phrases related to safety and safe spaces

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

issance

-0.92

Lenin

-0.74

 disproportion

-0.73

 intensify

-0.73

ithing

-0.67

acio

-0.67

ribune

-0.65

ennial

-0.65

 nostalg

-0.64

 favor

-0.64

POSITIVE LOGITS

 safe

0.88

 perimeter

0.77

 safer

0.77

 Safe

0.77

 precautions

0.76

 safest

0.75

Safe

0.74

 havens

0.73

 unprotected

0.73

Saf

0.72

Activations Density 0.077%

phrases related to safety and safe spaces

No Comments

No Known Activations