INDEX

Explanations

important or notable pieces of information from news articles

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

 forgetting

-0.74

âĶģ

-0.68

 somet

-0.65

 bothering

-0.64

(>

-0.64

âĶĢâĶĢâĶĢâĶĢ

-0.63

worldly

-0.63

Especially

-0.63

elf

-0.62

SPA

-0.61

POSITIVE LOGITS

 consists

0.85

 consisted

0.84

 alleges

0.83

 underscores

0.82

 allows

0.82

 illustrates

0.81

 comprises

0.81

 represents

0.79

 reinforces

0.77

 describes

0.77

Activations Density 0.162%

important or notable pieces of information from news articles

No Comments

No Known Activations