INDEX

Explanations

language-related terms and skills, particularly focusing on communication, speech, and written language

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

roxy

-0.87

romeda

-0.76

Ara

-0.71

ramid

-0.71

 Skydragon

-0.70

oppable

-0.67

kefeller

-0.67

 Profit

-0.67

MFT

-0.67

ONSORED

-0.66

POSITIVE LOGITS

 spoken

1.35

language

1.22

 phrases

1.22

 uttered

1.21

mith

1.16

 language

1.12

 interpreter

1.10

 translation

1.07

 words

1.06

 translations

1.02

Activations Density 2.457%

language-related terms and skills, particularly focusing on communication, speech, and written language

No Comments

No Known Activations