INDEX

Explanations

dates or durations

oai_token-act-pair · gpt-3.5-turbo

temporal phrases indicating ranges or durations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

airy

-0.76

aic

-0.72

>>\

-0.68

ETF

-0.67

Article

-0.67

kai

-0.62

ister

-0.60

tek

-0.60

Reloaded

-0.60

ECD

-0.59

POSITIVE LOGITS

cliffe

0.75

thood

0.74

essee

0.73

 mosqu

0.67

ved

0.66

termination

0.64

ocaly

0.62

quartered

0.60

osuke

0.59

essler

0.59

Activations Density 0.047%

dates or durations

temporal phrases indicating ranges or durations

No Comments

No Known Activations