INDEX

Explanations

conjunctions and relational phrases indicating connection or involvement

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

dfx

-0.76

dq

-0.67

æŃ¦

-0.65

Standing

-0.65

Abstract

-0.64

offensive

-0.62

Fal

-0.62

mA

-0.62

othal

-0.61

Hum

-0.61

POSITIVE LOGITS

 subscribe

1.13

romeda

1.09

 donate

0.82

rew

0.81

 enjoy

0.80

rea

0.79

istration

0.76

 then

0.75

 receive

0.75

 download

0.75

Activations Density 0.056%

conjunctions and relational phrases indicating connection or involvement

No Comments

No Known Activations