INDEX

Explanations

sentences where a statement is made followed by a subsequent clarification or additional information

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate contradictory or nuanced statements

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

hiba

-0.77

 sidx

-0.73

 sqor

-0.69

everal

-0.66

 Cosponsors

-0.65

aminer

-0.64

roxy

-0.64

 Uriel

-0.62

perty

-0.61

 Quote

-0.60

POSITIVE LOGITS

 necessarily

1.22

 anymore

1.04

 anything

1.00

nor

0.99

nor

0.94

 magically

0.90

any

0.88

ANY

0.80

 infall

0.79

anything

0.79

Activations Density 0.299%

sentences where a statement is made followed by a subsequent clarification or additional information

phrases that indicate contradictory or nuanced statements

No Comments

No Known Activations