INDEX

Explanations

punctuation marks, particularly questions and exclamations

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

sels

-0.76

ãĥ«

-0.69

ancies

-0.69

çĶ°

-0.67

éĹĺ

-0.66

ãĥĲ

-0.65

 misunder

-0.64

imar

-0.64

 recl

-0.63

 banned

-0.63

POSITIVE LOGITS

 Then

1.02

 Because

0.96

 Luckily

0.93

/"

0.92

 Obviously

0.90

 Nobody

0.90

 Knowing

0.89

 Sometimes

0.88

 That

0.88

 Which

0.85

Activations Density 0.051%

punctuation marks, particularly questions and exclamations

No Comments

No Known Activations