INDEX

Explanations

statements expressing repeated opinions or claims

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

rift

-0.78

allery

-0.68

elf

-0.68

 Gallery

-0.66

oreal

-0.64

swick

-0.63

 downtime

-0.62

 resil

-0.62

 fragmentation

-0.62

TOR

-0.61

POSITIVE LOGITS

 aloud

1.12

 loud

1.00

 uttered

0.98

 loudly

0.98

 echoed

0.97

 louder

0.97

 sounded

0.90

 reiterated

0.89

 whispered

0.85

 eloqu

0.81

Activations Density 0.203%

statements expressing repeated opinions or claims

No Comments

No Known Activations