INDEX

Explanations

references to 'truth' and related concepts of honesty and transparency

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 22-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L22R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.22.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Hij

-0.15

ICLE

-0.15

iac

-0.14

equal

-0.14

haus

-0.14

hire

-0.14

zet

-0.14

à¹Ħà¸§

-0.14

herited

-0.13

hof

-0.13

POSITIVE LOGITS

fulness

0.35

fully

0.34

iness

0.28

FUL

0.22

 serum

0.22

FULL

0.21

 Serum

0.21

full

0.20

worthy

0.19

ilde

0.18

Activations Density 0.015%

references to 'truth' and related concepts of honesty and transparency

No Comments

No Known Activations