INDEX

Explanations

time-related phrases, particularly durations and timelines

oai_token-act-pair · gpt-3.5-turbo

durations of time or periods taken for processes or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ħĭ

-0.96

rored

-0.72

ntil

-0.70

ranch

-0.68

 referen

-0.66

oci

-0.66

haps

-0.65

enture

-0.64

²

-0.64

riad

-0.64

POSITIVE LOGITS

 apiece

0.87

 longer

0.86

(~

0.84

 elapsed

0.82

 before

0.75

to

0.72

 convincing

0.70

 consecut

0.67

 patiently

0.64

 depending

0.64

Activations Density 0.079%

time-related phrases, particularly durations and timelines

durations of time or periods taken for processes or events

No Comments

No Known Activations