INDEX

Explanations

references to specific locations

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

Magikarp

-0.70

VD

-0.68

Phase

-0.68

creen

-0.65

Reloaded

-0.64

 reckoning

-0.63

jad

-0.62

 Grounds

-0.61

mare

-0.59

spect

-0.59

POSITIVE LOGITS

ality

1.17

alities

1.08

neys

1.02

agall

0.94

als

0.93

folio

0.93

 Clinton

0.89

ney

0.84

ilage

0.84

 Rodham

0.82

Activations Density 3.220%

references to specific locations

No Comments

No Known Activations