INDEX

Explanations

words related to specific characters or figures, particularly in a narrative context

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

DOS

-0.82

 Disp

-0.73

 Cafe

-0.73

 FedEx

-0.72

 Nicaragua

-0.71

ë

-0.70

ãģ¦

-0.70

Dan

-0.70

ãĤĭ

-0.69

EXP

-0.69

POSITIVE LOGITS

ou

1.43

roth

1.35

oul

1.25

ath

1.16

thal

1.13

oth

1.12

th

1.11

oup

1.09

arth

1.08

oun

1.07

Activations Density 0.047%

words related to specific characters or figures, particularly in a narrative context

No Comments

No Known Activations