INDEX

Explanations

dialogues or interview transcripts

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

 synerg

-0.70

 downstream

-0.68

 swast

-0.66

 harms

-0.63

 agon

-0.62

 fanc

-0.62

 casc

-0.61

 discarded

-0.60

 unchecked

-0.60

 forgotten

-0.59

POSITIVE LOGITS

Answer

1.44

Interview

1.26

Yeah

1.23

Question

1.09

Absolutely

1.02

laughs

1.01

Narr

0.97

Well

0.96

 Answer

0.95

ANS

0.94

Activations Density 2.315%

dialogues or interview transcripts

No Comments

No Known Activations