INDEX

Explanations

descriptions of a wide range of things or topics

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

illard

-0.68

nces

-0.68

LV

-0.66

 Keeper

-0.65

LU

-0.64

 Maid

-0.63

erick

-0.61

 Patron

-0.60

icks

-0.59

 Uriel

-0.59

POSITIVE LOGITS

 spectrum

1.11

sheet

1.10

sheets

1.09

 swath

1.04

 scope

0.99

spread

0.98

ranging

0.97

creen

0.97

 ranging

0.93

leaf

0.90

Activations Density 0.578%

descriptions of a wide range of things or topics

No Comments

No Known Activations