INDEX

Explanations

expressions of worthiness or value

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

 Frazier

-0.71

Zig

-0.70

eteria

-0.69

gdala

-0.67

ulated

-0.65

 Leilan

-0.64

acid

-0.63

ole

-0.62

esville

-0.61

ulating

-0.61

POSITIVE LOGITS

minded

0.84

 applause

0.80

 consideration

0.78

 deserving

0.77

 contenders

0.75

ãģĨ

0.73

 aspirations

0.71

 worthy

0.71

OGR

0.71

 successor

0.71

Activations Density 0.007%

expressions of worthiness or value

No Comments

No Known Activations