INDEX

Explanations

technology and business-related terms, potentially focusing on social media and restaurant collaborations

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

¿½

-0.88

 Howe

-0.68

 Marketable

-0.66

 Corm

-0.63

 handwritten

-0.61

Ļ½

-0.61

Ri

-0.61

 Revis

-0.61

è£

-0.59

©¶æ¥µ

-0.59

POSITIVE LOGITS

ups

1.00

up

0.98

based

0.88

art

0.87

out

0.86

ream

0.85

outs

0.83

ulates

0.82

rush

0.82

backed

0.81

Activations Density 0.043%

technology and business-related terms, potentially focusing on social media and restaurant collaborations

No Comments

No Known Activations