INDEX

Explanations

words related to achievements or accomplishments

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

PsyNetMessage

-0.90

EMBER

-0.69

ãĤ£

-0.69

ãĥ´ãĤ¡

-0.68

oding

-0.68

orescence

-0.68

BuyableInstoreAndOnline

-0.67

 unfocusedRange

-0.67

ãģ®å

-0.66

ATIONAL

-0.66

POSITIVE LOGITS

ety

1.13

leness

1.13

ery

1.12

erman

1.12

ering

1.06

ett

1.05

lers

1.02

eters

1.00

ening

0.99

eter

0.99

Activations Density 0.081%

words related to achievements or accomplishments

No Comments

No Known Activations