INDEX

Explanations

instances of the word "ast," indicating a focus on negative or harsh events

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

BOOK

-0.67

Ib

-0.63

BY

-0.59

 Sochi

-0.58

 Forth

-0.57

 Finder

-0.56

CTR

-0.56

pheus

-0.55

 largeDownload

-0.55

BILITIES

-0.55

POSITIVE LOGITS

eful

1.08

ening

0.92

liest

0.91

ened

0.91

liness

0.90

ener

0.86

eland

0.84

iest

0.82

eners

0.81

est

0.80

Activations Density 0.009%

instances of the word "ast," indicating a focus on negative or harsh events

No Comments

No Known Activations