INDEX

Explanations

phrases related to news stories

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Story" followed by numerical indicators, suggesting a focus on narrative structure or references to ongoing content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 stocking

-0.85

orpor

-0.85

İĭ

-0.74

ensor

-0.73

missible

-0.72

 congreg

-0.71

clusively

-0.71

 duty

-0.71

ikini

-0.71

itionally

-0.70

POSITIVE LOGITS

 Story

0.99

Story

0.95

telling

0.91

 Continued

0.90

 Stories

0.90

 Transcript

0.87

story

0.84

Edited

0.83

bold

0.82

Maker

0.77

Activations Density 0.009%

phrases related to news stories

instances of the word "Story" followed by numerical indicators, suggesting a focus on narrative structure or references to ongoing content

No Comments

No Known Activations