INDEX

Explanations

words related to reading, skipping, and textual analysis

oai_token-act-pair · gpt-3.5-turbo

references to complex narrative structures or themes in storytelling

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Pwr

-0.52

 Maintenance

-0.48

Opt

-0.47

)"

-0.46

entary

-0.45

 Cellular

-0.44

 Bang

-0.44

 Blend

-0.44

 Ended

-0.43

 Soft

-0.43

POSITIVE LOGITS

 earnest

0.51

 denomin

0.50

 sensibilities

0.49

 philosoph

0.49

rir

0.49

 ideological

0.48

 downright

0.48

 unch

0.47

 neocons

0.47

 decidedly

0.47

Activations Density 2.853%

words related to reading, skipping, and textual analysis

references to complex narrative structures or themes in storytelling

No Comments

No Known Activations