INDEX

Explanations

conjunctions and phrases indicating conditions or contrasts

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

ãĤ´ãĥ³

-0.73

 Â®

-0.67

zu

-0.65

!:

-0.62

hered

-0.60

aily

-0.60

ãĥĭ

-0.59

antine

-0.59

iren

-0.58

iden

-0.58

POSITIVE LOGITS

 blah

1.11

 stuff

1.11

 everybody

1.06

romeda

1.00

 hopefully

0.98

 yeah

0.97

 maybe

0.94

 secondly

0.93

 frankly

0.92

 basically

0.89

Activations Density 0.187%

conjunctions and phrases indicating conditions or contrasts

No Comments

No Known Activations