INDEX

Explanations

statistics or numerical quantities that have increased or accumulated over time

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "so" in various contexts

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

theless

-0.79

 realities

-0.63

 Arch

-0.58

amb

-0.56

 arch

-0.54

 arcs

-0.54

 wiser

-0.54

rals

-0.54

 Halls

-0.54

 gallery

-0.53

POSITIVE LOGITS

othes

1.14

oths

1.09

far

1.05

bered

1.04

iled

1.02

far

1.01

othe

0.96

apy

0.91

othing

0.90

aring

0.89

Activations Density 0.097%

statistics or numerical quantities that have increased or accumulated over time

occurrences of the word "so" in various contexts

No Comments

No Known Activations