INDEX

Explanations

phrases indicating a desire or action to put an end to something

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Digest

-0.68

etheless

-0.62

 immensely

-0.58

days

-0.58

 translations

-0.54

 Cavern

-0.54

<-

-0.53

 NETWORK

-0.52

quire

-0.52

 Provides

-0.52

POSITIVE LOGITS

 aside

0.98

 onto

0.90

lid

0.86

 brakes

0.85

 spotlight

0.82

 torch

0.82

 groundwork

0.74

 reins

0.74

 smack

0.72

 emphasis

0.72

Activations Density 2.490%

phrases indicating a desire or action to put an end to something

No Comments

No Known Activations