INDEX

Explanations

negative assessments related to personal circumstances or interactions

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

-prepend

-0.18

 beyond

-0.16

 indeed

-0.15

inde

-0.15

initely

-0.15

lah

-0.15

 oltre

-0.15

rego

-0.14

 Beyond

-0.14

_helpers

-0.14

POSITIVE LOGITS

 simply

0.20

Unless

0.19

 unless

0.19

 Unless

0.18

unless

0.17

 absent

0.16

 even

0.16

 early

0.15

 Simply

0.15

 EVEN

0.15

Activations Density 0.032%

negative assessments related to personal circumstances or interactions

No Known Activations