INDEX

Explanations

instances where the text references events or content that occurred in the past

oai_token-act-pair · gpt-3.5-turbo

repeated words or phrases indicating time periods or historical references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

"]=>

-0.66

rano

-0.66

PF

-0.66

Like

-0.66

Specific

-0.66

Website

-0.65

MON

-0.64

MSN

-0.64

ello

-0.64

IUM

-0.64

POSITIVE LOGITS

 captivity

0.81

order

0.81

 caves

0.72

roads

0.71

umbered

0.69

 Rampage

0.68

clusively

0.68

 earthqu

0.68

animate

0.67

sight

0.66

Activations Density 0.049%

instances where the text references events or content that occurred in the past

repeated words or phrases indicating time periods or historical references

No Comments

No Known Activations