INDEX

Explanations

words related to mythology and ancient legends, especially concerning entrances to realms like the underworld

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

pload

-0.79

 informants

-0.74

 informant

-0.72

 interceptions

-0.65

 interviews

-0.64

 Lauder

-0.64

 receipts

-0.63

 showers

-0.63

 inventory

-0.63

 clips

-0.63

POSITIVE LOGITS

themed

1.22

oriented

1.21

based

1.16

inspired

1.13

sized

1.11

focused

1.05

style

1.05

derived

1.04

centric

1.02

level

1.00

Activations Density 0.082%

words related to mythology and ancient legends, especially concerning entrances to realms like the underworld

No Comments

No Known Activations