INDEX

Explanations

words related to locations and titles

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

risk

-0.71

skill

-0.69

APS

-0.63

KEY

-0.62

wake

-0.61

interstitial

-0.60

pu

-0.60

ONEY

-0.60

ãĥ´

-0.60

plays

-0.59

POSITIVE LOGITS

llor

0.89

osterone

0.79

riors

0.77

terday

0.75

ificate

0.75

 hybrids

0.74

quished

0.70

tenance

0.70

arthed

0.69

bia

0.66

Activations Density 0.320%

words related to locations and titles

No Comments

No Known Activations