INDEX

Explanations

terms related to conviction or persuasion through detailed storytelling or narration

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Seym

-0.85

ropolitan

-0.77

erity

-0.74

ngth

-0.69

 Dwell

-0.65

 Alibaba

-0.64

lain

-0.61

 deductible

-0.60

deen

-0.60

ablishment

-0.57

POSITIVE LOGITS

ces

0.99

ci

0.89

osaurs

0.89

ius

0.88

icum

0.87

istic

0.86

eteen

0.85

itus

0.85

ist

0.85

warm

0.83

Activations Density 0.036%

terms related to conviction or persuasion through detailed storytelling or narration

No Comments

No Known Activations