INDEX

Explanations

locations or settings within a narrative

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

icer

-0.82

issance

-0.81

icism

-0.74

olyn

-0.73

quel

-0.72

vous

-0.71

icial

-0.66

jin

-0.65

unes

-0.65

ework

-0.64

POSITIVE LOGITS

holders

1.17

holder

1.06

bos

1.04

 where

0.86

 locations

0.86

 holder

0.84

abouts

0.78

lot

0.78

wic

0.77

 frequ

0.77

Activations Density 0.802%

locations or settings within a narrative

No Comments

No Known Activations