INDEX

Explanations

named locations combined with educational or professional entities

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 cache

-0.63

 hoard

-0.61

ulkan

-0.60

 expans

-0.59

è¦ļéĨĴ

-0.59

 Bots

-0.58

XY

-0.58

 commercially

-0.57

 cockpit

-0.57

 nesting

-0.56

POSITIVE LOGITS

uates

1.39

 Students

1.34

 tuition

1.32

Students

1.31

 faculty

1.30

 Faculty

1.29

 students

1.29

student

1.28

 graduate

1.28

 campuses

1.26

Activations Density 3.188%

named locations combined with educational or professional entities

No Comments

No Known Activations