INDEX

Explanations

references to the "first" occurrence of events or elements

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

tics

-0.82

mbuds

-0.81

athed

-0.66

borg

-0.66

aths

-0.65

md

-0.65

ITH

-0.64

sav

-0.64

gregation

-0.64

rs

-0.64

POSITIVE LOGITS

 iteration

1.06

 installment

1.04

 batch

0.99

 couple

0.98

 thing

0.98

few

0.97

 baseman

0.93

 responders

0.92

 volley

0.92

 lady

0.89

Activations Density 0.067%

references to the "first" occurrence of events or elements

No Comments

No Known Activations