INDEX

Explanations

terms related to a series of connected or interlinked items

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "chain."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.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

 Vern

-0.70

 Lowell

-0.68

PP

-0.66

fol

-0.66

humans

-0.65

essional

-0.65

ukong

-0.62

hett

-0.62

 Glenn

-0.62

glas

-0.62

POSITIVE LOGITS

 chain

4.12

 chains

3.14

 Chain

2.81

chain

2.70

Chain

2.50

chains

2.08

 Chains

2.00

 chained

1.42

 cascade

1.33

 hierarchy

1.00

Activations Density 0.006%

terms related to a series of connected or interlinked items

occurrences of the word "chain."

No Comments

No Known Activations