INDEX

Explanations

numerical scores or statistical data relevant to performance

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Macy

-0.15

ngth

-0.13

pch

-0.13

|_

-0.13

utos

-0.13

astes

-0.13

reon

-0.13

ila

-0.13

atham

-0.13

ILA

-0.13

POSITIVE LOGITS

lund

0.16

PROTO

0.14

ÙĥÙĬØ©

0.14

ãĥ¼ãĤ¸

0.14

sled

0.13

otel

0.13

.self

0.13

Ping

0.13

/--

0.13

 unary

0.13

Activations Density 0.016%

numerical scores or statistical data relevant to performance

No Comments

No Known Activations