INDEX

Explanations

verbs expressing intention or planned actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Solitaire

-0.75

acks

-0.66

adj

-0.64

anca

-0.61

EStreamFrame

-0.60

Americ

-0.60

 archetype

-0.58

java

-0.58

lc

-0.57

unch

-0.57

POSITIVE LOGITS

lessly

0.79

fully

0.72

 provoking

0.70

phis

0.69

orial

0.69

untarily

0.69

 Parenthood

0.68

ller

0.68

eering

0.68

atively

0.68

Activations Density 0.017%

verbs expressing intention or planned actions

No Comments

No Known Activations