INDEX

Explanations

phrases related to depth, profoundness, or severity

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

Ĥİ

-0.80

haw

-0.69

amazon

-0.67

ordan

-0.66

essors

-0.64

ATT

-0.61

ãĥ¼ãĥ³

-0.60

uthor

-0.60

ivas

-0.60

EED

-0.59

POSITIVE LOGITS

 vein

1.11

 deep

0.97

 deeper

0.96

 trenches

0.95

 depths

0.93

dug

0.92

deep

0.92

 deepest

0.89

 depth

0.88

 buried

0.84

Activations Density 2.431%

phrases related to depth, profoundness, or severity

No Comments

No Known Activations