INDEX

Explanations

phrases related to analysis and speculation about given situations

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

jer

-0.91

ereo

-0.86

scribe

-0.82

rones

-0.74

inse

-0.74

ilyn

-0.71

Ã©s

-0.69

okes

-0.68

eware

-0.68

inis

-0.68

POSITIVE LOGITS

 circumstances

1.04

 hindsight

0.83

 precedent

0.81

 ample

0.79

 sheer

0.77

how

0.76

 inexper

0.74

 precedence

0.74

 appearances

0.73

 context

0.73

Activations Density 0.566%

phrases related to analysis and speculation about given situations

No Comments

No Known Activations