INDEX

Explanations

phrases that indicate preference or favor towards something

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

awar

-0.83

block

-0.75

/-

-0.69

}}

-0.69

UGH

-0.65

break

-0.65

stan

-0.65

wow

-0.64

nor

-0.62

eno

-0.62

POSITIVE LOGITS

 simpler

0.87

 simplified

0.82

 embracing

0.77

 softer

0.76

 streamlined

0.74

 something

0.71

 sleek

0.70

 trendy

0.70

 concentrating

0.69

 embrace

0.67

Activations Density 0.075%

phrases that indicate preference or favor towards something

No Comments

No Known Activations