INDEX

Explanations

repeated phrases or elements indicating emphasis and intensity

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

SPONSORED

-1.04

elson

-0.83

enced

-0.82

yll

-0.81

edit

-0.78

uin

-0.78

thood

-0.77

nir

-0.76

erson

-0.75

isu

-0.73

POSITIVE LOGITS

 slack

0.89

 basics

0.88

 stink

0.86

 courage

0.85

 stairs

0.81

 seams

0.80

 pace

0.79

 defenses

0.78

 playbook

0.76

 ante

0.76

Activations Density 0.119%

repeated phrases or elements indicating emphasis and intensity

No Comments

No Known Activations