INDEX

Explanations

statements that introduce or follow a conditional scenario

oai_token-act-pair · gpt-3.5-turbo

mentions of historical events, rumors, and questions of validity or trustworthiness in reports and claims

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

involved

-0.71

({

-0.64

Rus

-0.61

ciating

-0.61

ILE

-0.61

icular

-0.60

ilaterally

-0.60

dos

-0.60

bat

-0.60

CTR

-0.60

POSITIVE LOGITS

 indeed

0.88

 correct

0.85

 accurate

0.85

 replicated

0.85

 true

0.84

true

0.76

 extrap

0.75

 anything

0.72

any

0.70

rition

0.69

Activations Density 0.134%

statements that introduce or follow a conditional scenario

mentions of historical events, rumors, and questions of validity or trustworthiness in reports and claims

No Comments

No Known Activations