INDEX

Explanations

names of friends or companions mentioned within several texts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etting

-0.76

ento

-0.68

itters

-0.67

ahime

-0.65

raltar

-0.65

cale

-0.65

ettings

-0.65

eway

-0.64

earchers

-0.64

 ceilings

-0.64

POSITIVE LOGITS

 Stephen

0.86

 Rupert

0.86

 Jaime

0.85

 Joseph

0.83

 Sidney

0.83

 Theo

0.83

 Nicolas

0.83

 Hannah

0.82

 Sophie

0.82

 Philip

0.81

Activations Density 0.130%

names of friends or companions mentioned within several texts

No Comments

No Known Activations