INDEX

Explanations

parts of conversations that involve apologies, requests for consideration, and responses indicating reluctance or concern

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

requent

-0.51

antium

-0.51

minent

-0.45

ionais

-0.44

탤

-0.44

adays

-0.43

prehensive

-0.43

 recombinant

-0.43

tools

-0.42

ousands

-0.42

POSITIVE LOGITS

BufferException

0.45

UseVisualStyle

0.42

PyExc

0.39

DebuggerStep

0.38

mmä

0.36

 spoke

0.36

 overheard

0.36

 speechless

0.36

 snippetHide

0.36

 straighten

0.35

Activations Density 0.012%

parts of conversations that involve apologies, requests for consideration, and responses indicating reluctance or concern

No Comments

No Known Activations