INDEX

Explanations

personal pronouns and verbs indicating actions or intent

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Consortium

-0.77

 Eleven

-0.64

 Binding

-0.64

 Globe

-0.61

 Amazing

-0.60

 Standard

-0.60

 Rousse

-0.59

 Associated

-0.59

 Anat

-0.59

 Wide

-0.58

POSITIVE LOGITS

've

1.51

're

1.40

'd

1.34

'll

1.25

'm

1.09

 knew

1.09

 could

1.05

 encount

1.05

can

1.03

 forgot

0.99

Activations Density 1.772%

personal pronouns and verbs indicating actions or intent

No Comments

No Known Activations