INDEX

Explanations

words related to time or ordering

oai_token-act-pair · gpt-3.5-turbo

occurrences of temporal markers and ordinal indicators related to events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

cause

-0.70

eker

-0.68

onto

-0.61

ourge

-0.61

oline

-0.59

||||

-0.59

kefeller

-0.58

ctrl

-0.58

 outweigh

-0.58

adeon

-0.57

POSITIVE LOGITS

 interviews

1.07

 interview

1.01

 instance

0.98

 cases

0.95

 nutshell

0.94

 case

0.93

 instances

0.89

 context

0.87

 excerpt

0.85

 example

0.84

Activations Density 0.235%

words related to time or ordering

occurrences of temporal markers and ordinal indicators related to events

No Comments

No Known Activations