INDEX

Explanations

numbers with specific formats, specifically those formatted as 4-3-7 or similar

oai_token-act-pair · gpt-3.5-turbo

patterns associated with numerical sequences or statistics

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

DERR

-0.72

 cutting

-0.69

 Cutting

-0.65

 quitting

-0.64

 Paddock

-0.62

 exit

-0.62

 core

-0.61

 exiting

-0.60

 casting

-0.60

 scream

-0.59

POSITIVE LOGITS

©¶æ

1.02

cent

0.84

cell

0.82

infect

0.82

distance

0.81

inf

0.81

whatever

0.81

edition

0.80

tion

0.80

turned

0.79

Activations Density 0.097%

numbers with specific formats, specifically those formatted as 4-3-7 or similar

patterns associated with numerical sequences or statistics

No Comments

No Known Activations