INDEX

Explanations

instances of the word "time" with relatively high activation values

oai_token-act-pair · gpt-3.5-turbo

mentions of time and temporal references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ramid

-0.76

ouk

-0.70

RD

-0.69

Bey

-0.69

iants

-0.63

reditary

-0.62

atl

-0.61

MAT

-0.60

rim

-0.60

atever

-0.59

POSITIVE LOGITS

frames

0.89

 elapsed

0.86

zone

0.85

 interval

0.79

 pause

0.76

 consuming

0.70

ago

0.69

 msec

0.69

 traveller

0.69

][

0.69

Activations Density 0.026%

instances of the word "time" with relatively high activation values

mentions of time and temporal references

No Comments

No Known Activations