INDEX

Explanations

medical terms and drugs

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

natureconservancy

-0.65

xtap

-0.64

conn

-0.63

ĵĺ

-0.63

=-=-=-=-=-=-=-=-

-0.61

Scope

-0.60

essim

-0.59

ModLoader

-0.58

ĻĤ

-0.58

ãĤ©

-0.57

POSITIVE LOGITS

forth

0.82

rimination

0.67

warts

0.64

lict

0.64

 Forth

0.63

oche

0.63

adder

0.63

 Warwick

0.63

hyde

0.63

arb

0.62

Activations Density 5.301%

medical terms and drugs

No Comments

No Known Activations