INDEX

Explanations

sentences related to discussing the methodology or assumptions of logical reasoning and academic arguments

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

 Boot

-0.64

 Inher

-0.64

dylib

-0.63

 hail

-0.61

 Strikes

-0.60

Jur

-0.60

 Ital

-0.60

ebted

-0.60

Kut

-0.59

 watches

-0.59

POSITIVE LOGITS

 preferable

1.11

 counterproductive

1.11

 fraught

1.08

 advisable

0.99

 frowned

0.95

 futile

0.95

 problematic

0.94

 impractical

0.94

 folly

0.91

 daunting

0.90

Activations Density 2.503%

sentences related to discussing the methodology or assumptions of logical reasoning and academic arguments

No Comments

No Known Activations